Sharing my tune for feedback: GB60, GB25, GB10. DSP408

[Julian]

I put all those in REW and tweaked a bit more. Some of the REW crossovers do not exists in the DSP408 so some tweaking was needed. It doesnt measure perfectly, but it sounds amazing. There are a couple frequencies that are off but I cant tell with my ears.
I was listening to this. It is really well recorded.
https://youtu.be/5ewK2_-IRTg

I am going to enjoy it as it is right now and tweak it later as needed.

[Jdunk54nl]

You should have bessel 18db/Oct (be3) and linkwitz 24 db/Oct (lr4)

Also, after using only REW, I always have to make minor adjustments. Some of the predictions just don't work as well as rew thinks they will.

[Julian]

I should have looked up what the filters mean in REW. I incorrectly assumed the meaning of the numbers after the filter (4 in lr4). And my mind could not grasp that a $180 DSP could do everything. Doh!

I am still impressed with how accurate REW was. I suppose, through acoustic wave superposition, changing filters affects how the waves sum at other frequencies.

[Jdunk54nl]

Here was the drivers side:

You need to drop this by 1db. It was 1db hotter than the passenger side when EQ'ing.

[Jdunk54nl]

@Justin or @dumdum,

Can either of you talk about the phase differences on the low end between these two? Is it going to be noticeable and change things? I haven't investigated it enough to know this answer:

[Julian]

Here was the drivers side:

You need to drop this by 1db. It was 1db hotter than the passenger side when EQ'ing.

Thanks! I tried to account for that. What I did, and please let me know if this is wrong, is observe the level of the target curve in REW for each driver, note the differences, and correct the gain in the DSP accordingly.

I will then confirm at the end with an actual measurement using the SPL meter and final values in the response curves.

[Justin Zazzi]

@Justin or @dumdum,
Can either of you talk about the phase differences on the low end between these two? Is it going to be noticeable and change things? I haven't investigated it enough to know this answer:

I'm not sure, phase response is not easy to predict at a glance. You can also consider the phase response of the other speaker that you are matching it to. So if the midbass in the door has a complimentary phase response to one of those plots, then it will sum together more better.

This is a bit like how electrical filters don't matter and only the acoustic response does: in that the absolute phase response of one speaker doesn't matter too much as long as it is a complimentary match to the next one in the crossover. This leads back to the inception of my tuning spreadsheet: by creating targets and trying to get a specific acoustic response, we have a good chance of creating that complimentary pair of phase responses. Curveception?

[Julian]

I have been doing more tuning with some different measurement methods. Thanks to Jdunk for pointing this out. Before, I was sitting in the seat and moving the mic around my ears, switching sides, mic aimed forward. I was getting inconsistent measurements and some substantial differences between the left and right ears.
So I reclined my seat to get my self out of the zone of measurement, did circles around where my head would be and used a 90 degree cal file with the mic pointed vertically.

I got much more rational results and it was also a lot more consistent.
The 3rd one is both together. Noise floor is a bit different in these as I had the truck running for some of the measurements.

I am still struggling with making efficient use of the 10 filters in REW to get a better match. Getting the crossovers correct sure helps, but I think spending more time on each individual filter and thinking about how they react is also helping.

Here are my original REW responses for Left/Right.

https://drive.google.com/file/d/1CM6...ew?usp=sharing
https://drive.google.com/file/d/1sGZ...ew?usp=sharing

Let me know how I am doing please. I am going to keep at it trying for better results each time. Gives me something to do while I wait for my tweeter sails to be finished.

I am still getting some weird issues at narrow frequencies as well. Hoping a closer match sorts that out.

[Jdunk54nl]

Looks good overall. Your low pass filter out, slope looks right, but might need to reduce the cut off frequency.

[jtrosky]

Were they all taken at the same volume level? Just curious why the level isn't any higher on the "both" measurement (unless you took the combined measurement at a lower volume level?)...

I just want to make sure that you are getting proper summation when both L+R are playing together. I would expect the L+R combined measurement to be 4dB to 6dB higher than the individual measurements.